Pharmaceutical companies spend billions of dollars annually on one-to-one promotion of their products to physicians. The most common method is that of a sales representative calling on physicians' offices to engage them in product discussions with the use of carefully prepared promotional materials, known as a “detail” within the industry. Pharmaceutical companies devote considerable resources to creating and testing these materials and in preparing representatives to effectively use them. The total detail—promotional materials and voiceover—is not evaluated in a consistent manner such that findings may be compared across products and markets, and used to predict the impact of a detail on physician behavior and prescribing. As a result, companies do not isolate and promulgate best practices, nor can they fully assess the effectiveness of the campaign prior to field implementation.
Promotional materials are the items that sales representatives put in front of a customer during a detail to gain their attention, convey marketing messages, or prove a claimed benefit. When accompanied with the salesperson's commentary, the materials are designed to make the presentation more effective and more memorable, thereby increasing the likelihood that the customer will take the desired action. These materials may take a variety of forms, from a one page sell sheet or dosing reminder card to pamphlets, sample boxes, and clinical study reprints. Once created and approved, these materials may be used for weeks to several years. Effective approaches and methods may be extended to other product teams within the organization or revived years later.
In constructing the materials, marketers must strive to distill complex information into extremely concise messages that will influence physicians, and to do so within FDA guidelines. To effectively use the materials in a detail, they pull their sales forces from the field two to three times a year to explain the data and rehearse the voiceover. In testing the materials, marketing teams leave their core duties to travel to cities throughout the US for focus group and other qualitative testing.
Companies use a variety of methods to test detailing. This testing is designed to ensure that the promotional materials faithfully and effectively convey the brand's intended messaging, that the materials are accurate, intelligible and graphically appealing and that, ultimately, the total detail helps to position the brand in a way to increase usage. At present, this testing is predominately focused on the physical materials, although occasionally researchers will add a simulated sales pitch to the presentation of the materials to better simulate the selling environment.
Researchers make a series of methodologic choices that determine the ultimate validity of the findings and—by extension—the ability of the research to predict success of the campaign in the marketplace. The major methodologic components of marketing research include the type of interview, the sample, the survey instrument and the analysis, among other components.
The first decision for a researcher is the type of interview: qualitative or quantitative. At present, qualitative testing is the predominate method for testing details in the pharmaceutical industry, mainly due to tradition and the absence of other options. Each interview type has a role in marketing research; a brief discussion of the method, its limitations and broader applicability for promotion research follows.
Qualitative testing: When a subject is poorly understood or the researcher needs license to reactively delve into particular responses, qualitative research is preferred. An example of this technique is a fill-in-the-blank, or open-ended, question. This method can produce a deep, anecdotal understanding of the subject area among those who can physically witness the research; this understanding is accurate to the extent that other methodologic choices are adequately controlled. This form of research is conducted with a small number of respondents, either as one-on-one interviews or as part of a focus group. These interviews are typically conducted by a trained moderator who guides the participants through the materials and provokes a reaction. In qualitative testing, the respondent receives some guidance in allowable responses but is, for the most part, answering a series of open-ended questions.
Most often, these interactions are scripted in the form of a discussion guide, although the moderator may frequently deviate from this script to react to an individual respondent's opinions. The presence of the interviewer imparts a subtle bias on the results—it is instinctively more difficult to be negative in front of another person, even if the researcher disavows any connection with the marketer. Responses are captured in hand-written notes or video/audio taped for later transcription, and may be produced as a rough grouping of verbatim responses or as a list of those responses. Rarely, these results are coded into a quantitative report.
Qualitative testing places considerable limitations on the interpretation of results. With respect to the adequacy of the sample, qualitative testing of large groups is prohibitively expensive and time consuming. As a result, samples are typically too small to analyze by subgroup and too unrepresentative and geographically concentrated to reliably project the findings to the entire physician universe. With respect to analysis, qualitative findings are typically reported without quantification, as the relatively unstructured method of the interview produces anecdotal data not suited for quantitative analysis. In this way, qualitative testing is much like a conversation, and as such tends to deviate from interaction to interaction. Although the ability to analyze results across campaigns and across time is patently helpful to the marketer, the inherent nature of qualitative results data defy attempts to benchmark a campaign against others or to gain full value of the data over time.
Another limitation of qualitative results data is the need for a marketer to attend the research to gain full benefit. In essence, hearing third-party anecdotes from a conversation is markedly different, and less valuable, than hearing the entire conversation. Because of the absence of coherent results from free-form data, a failure to attend can effectively remove an absent marketer from understanding and influencing the process. This situation results in marketers abandoning their core duties to travel to multiple cities to “sit in on” the research. As a result, marketers are resigned to the travel—and resultant diminution of productivity around other efforts—multiple times per year.
Quantitative testing: when a research area is sufficiently understood and when the researcher has a need to analyze findings for significance, by subgroup or across surveys, quantitative testing is preferred. Examples of this technique include multiple-choice, ranking and rating questions. This form of research exposes respondents to a consistent questionnaire via mail, fax, in-person interview, and more recently, via the Internet. Non-network variants (mail, fax, non-networked computer terminals, and in-person) provide a form or other structured question set to be filled out by the respondent or, in the case of an in-person method, by an interviewer; online variants render the questionnaire as software via a web browser. Online variations further increase the illusion of anonymity, thereby prompting more candid—and accurate—responses.
Non-personal fielding methods eliminate the biasing effect of a moderator and reduce testing variability, where the questionnaire might change slightly in each deployment. Internet deployments also allow the research to control for other environmental conditions, such as time available to take the survey, the ability to “read ahead” or “back up” when taking the survey, etc. As compared to mail or fax method, this Internet method allows the researcher to more carefully control the experiment, again affording the opportunity for more accurate findings.
With quantitative methods, researchers may choose to test small groups or to test thousands of interactions. Larger samples produce higher statistical confidence around findings and the ability to analyze across more subgroups but cost more and take more time; the smallest sample that will produce statistical significance is generally regarded to be no less than 30–40 per subgroup. Sample size is determined by modeling the research instrument against the desired resolving power of the survey.
Responses are recorded in a form designed for subsequent statistical testing and analysis, and are typically processed into cross-tabulated reports and/or by graphs and charts. Quantitative testing produces a highly analyzable, trendable data set but can miss a more nuanced understanding of responses due to the relative inability to reactively—non-programmatically—drill into responses.
In considering the other methodologic components, quantitative testing provides the largest safety net with respect to determining the believability of the response set and applying the findings in the marketplace. A quantitative sample can be large enough to provide statistical testing of difference between subgroups and can be representative of the larger population so the findings of a small group may be reliably projected to the entire population. The sample can also be geographically representative, diminishing the danger of localized or regional effects.
Although quantitative research offers the promise of a more consistent stimulus—through strict control of the survey instrument and the methodology—a desire to meet the specific needs of the customer may lead a researcher to tailor his methods to the needs of an individual marketing team. A survey instrument may thus become a customized instrument for each customer. While this customization might be minor, such as the inclusion of Reason G into a list of Reasons for Prescribing A–F, this customization may be sufficient to produce incomparable data. In this example, a ranking of Reason G as the highest rated attribute is meaningless when compared against data from respondents who were not afforded the same response choice. In this way, minor customization of the survey instrument can produce sufficient variability in survey stimuli to preclude analysis of what are, in essence, responses to different questions. In this fashion, the customization inherent in prior art methods are used in the pharmaceutical industry contributes to the lack of performance benchmarks of promotional materials across the industry.
Regardless of the type of interview, a significant limitation of current offerings—both qualitative and quantitative—is in the scope of testing. A sales detail is comprised of two main components, the promotional materials and the accompanying sales pitch. Present techniques tend to test these components in isolation or test only the visual materials, severely limiting a marketer's ability to accurately gauge the effect of the total detail. It is the combination of promotional materials and the accompanying voiceover that produces the intended effect in the marketplace—to test either in isolation is to test for a selling environment that does not exist. Results from this testing may impart an inaccurate view of how the campaign will perform in the marketplace.
Lastly, current methods do not provide an analysis of the strengths and weaknesses of the detail against industry norms, nor do they attempt to project the ultimate success of the campaign. First, there are no norms to compare against; second, rarely—if ever—do current methods provide detail around which discrete components are most effective; third, few methods calibrate their attitudinal findings with actual behavioral data.